#metabrainz

aerozol: please do not beat me over the head, that sounds painful

reosarevok: What’s the status of MBS-10603?

MBS-10603: Add Rhodesia to the list of release countries https://tickets.metabrainz.org/browse/MBS-10603

"If this needs to be separate, then you'll need to request it to be added as an actual separate area first"

(I expect drsaunde will say no and this will be closed)

It looks like an AREQ ticket had been opened beforehand, just not in the right way?

(It is linked to it.)

Yeah, but it was (correctly IMO) added as an alias

tbh, I'd just close this

I'll do that

It looks a bit like finding the right office in an administration building.

But the new outcome seems to make sense :)

"le laissez-passer A-38"?

re: The test earlier, I don't think that's it, the ::Edit test names the fields the same way

I think it might be an issue with unitialized artist credit not yet having the fields?

Possibly

reosarevok: Updated the docker-server-configs PR too.

I'm just not sure whether that means it just cannot be done with Perl alone and it has to be a selenium test or if there's a way I'm missing it

bitmap ^

zas: which version of docker should I downgrade to?

Lol @ le laissez-passer A-38, great reference

reosarevok: I came across this ticket as I was reviewing https://wiki.musicbrainz.org/Internationalizati...

Some points are likely relevant as tickets.

I assume there's no way to match those RFE links to anything current?

links are to an older bug tracker

Yeah. It's like two bug trackers out of date I think :D

SUHH now is in the database :)

"Some extra "countries" have been requested, notably for "Benelux", the European Union, "Unknown" and "Worldwide" releases RFE 987023." I think we have tickets for the first two, and obviously the last exists

I haven't heard of "Unknown" as a request I think?

But the wiki table for country codes is likely outdated, we should probably have MBS to generate it.

Well, we have https://musicbrainz.org/admin/attributes/Language but we don't show it to non-admins yet

(https://github.com/metabrainz/musicbrainz-serve...)

Might as well fix those conflicts while at it

mayhem: it only happens on docker upgrade, afaik, so, if you restarted containers, it shouldn't happen anymore until next upgrade. We reported it upstream, and it seems more people have same or similar issue, so I expect it to be fixed in next versions

i restarted containsers many many times, it keeps happening.

reosarevok: btw there are many conflicts in React conversion PRs

I can imagine

I'll look them over

downgrading to 5:23.0.6-1~ubuntu.20.04~focal

mayhem: then it's likely another issue, yes, you can downgrade <24

k

mayhem: does downgrading solve the issue for you?

everything started up again, but not enough time has passed for the problem to reoccur.

reosarevok: Re extra countries, they have all been requested and accepted but "Benelux" which has been refused as AREQ-2021 but that did not take the release country issue into account.

AREQ-2021: Benelux https://tickets.metabrainz.org/browse/AREQ-2021

Europe and Worldwide are currently available.

Europe is not the European Union, as users will remind each other of often

"Unknown Country" used to be available at least 8 years ago according to the documentation, probably been removed a while ago for a good reason.

That said, I think it makes sense not to add random groupings

The EU grows so an EU release from 1980 would be in a different area from an EU release in 2023

Oh right, Europe != EU.

And Benelux is just a grouping of three countries and it's not hard to just set the three if they apply IMO

For the release

Ok, it should probably be documented.

The other paragraphs about display are outdated.

lol, just realized I mixed up release countries and languages earlier

Sorry yvanzo

That said, I think the right place to list release countries is the same place as languages and whatnot, under /attributes

IIUC there's no way currently of editing the stuff in release_country from the website either?

If so, that should be part of the implementation I'd say :)

Countries are edited from area editing pages.

Those are area attributes.

It has nothing to do with types.

Gender is there, Language is there, it's not about types, it's about making stuff available to edit that is not elsewhere

Areas should be edited from the area pages, sure, but whether an area qualifies as a release country should be the decision of schema people, not area people, IMO, and indicated elsewhere

Gender is an attribute type, not an entity attribute.

Country codes are area attributes.

It’s available to edit already.

I see your point but there is no conflict with the area editor either :)

reosarevok: Where should country groups be documented?

I'm not talking about country codes, I'm talking about the release_country table that dictates which areas can be release countries.

Sorry, "country_area"

That's just a list of area ids, not related to country codes :)

You mean "worldwide" and "europe"? Honestly, not sure. Either on their respective annotations or under doc/Area I guess?

Alright I didn’t dig the details, thanks for having looked into it :)

reosarevok: I mean Benelux and stuff that won't be in the database but might appear on physical releases.

Oh

There is https://musicbrainz.org/doc/Release/Country now.

I guess that could be under /Release, possibly, saying that if it says it's for a specific set of countries you should set them

I just moved it from "Release Country".

Ok :) Maybe there then? :D

But I’m not sure why Europe is there and not Benelux?

Would probably fit under "What a Release Country is not"

Honestly, I'm not sure why Europe is there either

I think it just got grandfathered in

It kinda unknown but European.

Because it's fairly common to have "US release and European release" but it's a lot harder to figure out where in Europe it was available, I guess?

I expect that was the reasoning for this - it's kinda different from a Benelux release in that for that you know it applies to all three countries

But yes, it does basically seem to be "unknown, but some European countries, anyway" :p

aerozol for when you wake up: I suppose you made the background images for the recommendations page using a vector application? Could you please export them as SVGs? (these guys:) https://usercontent.irccloud-cdn.com/file/hd84V...

reosarevok: Up for review in the transclusion table

yvanzo: reworded a bit and transcluded

Thanks!

Updated the PR description in https://github.com/metabrainz/musicbrainz-serve... to not entirely suck, too

bitmap: I moved the postgres-health-check repo under metabrainz/ and added MIT license as discussed -> https://github.com/metabrainz/postgres-health-c...

<yvanzo> "Is the Tagger section...still up-to-date?" Not really. I'll take a look, but I think any outstanding issues are already captured on tickets.

yo akshaaatt

should i put unit test for API calls when the api returns 40X status codes too?

i guess yes

lucifer^^

jasje: yes

im pushing code to my draft pr

have a look

in a sec

will do later.

okeyy

zas: 👍

yvanzo: thanks for taking care of the docker-server-configs PRs. looking into reosarevok's RG test now

yvanzo, I checked with outsidecontext about the "Tagger" section of https://wiki.musicbrainz.org/Internationalization and his comment was, "That info is all totally outdated. There is some validity for the issue of the filesystem if non-Unicode encoding would be used. But in practice all modern OS deal well with the Unicode filenames Picard writes."

rdswift: Thanks, dropped

👍

bitmap: That reminds me that I initially wanted to update all working copies. On it!

oh good idea

Sure jasje

reosarevok: yeah it's likely because the AC fields are rendered by JS after the page loads. www::mechanize expects them to be in the HTML already

one way around this would be to just submit the post request manually instead of with the $mech->submit_form_ok api. let me see if I can find an example

though i'd recommend using a selenium test since you can interact with the fields on the page using JS

https://gist.github.com/mwiencek/d684d741cd174e... should work

bitmap: thanks, will try that in the morning. The goal of this is to have a reasonably standard set of tests for all Create controllers, and we already do for the other ones, hence wanting it here, although we could *also* have a more wide Selenium one added later :)

makes sense (though I'm not 100% sure if you can actually test the fix in the PR with the mechanize test?)

(guessing not since posting the form directly will bypass any template code entirely)

bitmap, reosarevok: Updated docker-server-configs and fixed their SSH setup (on all of MB hosts at least for now).

thanks! that will prevent some headaches in the future

zas: There was another glitch (fixed it for all hosts listed in docker-server-configs): /root/.ssh/known_hosts still had the old GitHub RSA key (on a few hosts only) or for some IPs (on most hosts).

So git operations would fail randomly depending on which IP was reached.

bitmap: yeah, I was wondering that myself, I might need to add a second Selenium test for that already now. Oh well. There's worse things than Selenium tests. Probably. Physical torture.

I'll look into it tomorrow, thanks :)

lol

if it's really that bad let me know how I can help and maybe we can improve things

It's just a lot more confusing to me than standard Perl or JS tests, but that's just me and I'd probably solve it by writing a lot of them so I get used to it really

should probably merge https://github.com/metabrainz/musicbrainz-serve...

It's just that the "writing a lot of them" part is not fun :p

I never tested that process yet. yvanzo: have you done that? ^

process of?

Writing selenium tests that use the search server, in order to know if that doc in the PR above is good :)

bitmap: https://github.com/metabrainz/musicbrainz-serve... seems ready to merge, btw, in any case

right, thanks

No was stuck with musicbrainz-docker.

Oh, I didn't mean recently, just like ever :D

Never got a working env for selenium.

Oh

Gotcha

[musicbrainz-server] 14mwiencek merged pull request #2951 (03master…mbs-13015): MBS-13015: Tracks with only URL relationships don't show in the relationship editor https://github.com/metabrainz/musicbrainz-serve...

Yeah, my selenium tests were a mess for a long time, I *think* they maybe pass now? Or maybe some still break

Usually I just wrote them, sent them up to CI for testing, hoped they pass, sigh

ok I should probably try setting it up with docker, I only ever did it on the host